3842
Biochemistry 1985, 24, 3842-3846
Thermal Denaturation of the Core Protein of lac Repressort Susan P. Manly, Kathleen S. Matthews, and Julian M. Sturtevant* Department of Biochemistry, Rice University, Houston, Texas 77251, and Departments of Chemistry and of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 0651 1 Received December 11, 1984
ABSTRACT: The thermal denaturation of the core protein of lac repressor was studied alone and in the presence of the inducer isopropyl 0-D-thiogalactoside (IPTG) and the antiinducer o-nitrophenyl P-D-fucoside (ONPF) by means of high-sensitivity differential scanning calorimetry. The denaturation that takes place a t about 65 OC is apparently irreversible; i.e., a rescan of a previously scanned sample of protein solution shows no denaturational endotherm. Despite this irreversibility, the denaturation appeared to follow quantitatively the dictates of equilibrium thermodynamics as embodied in the van? Hoff equation. The results obtained indicate clearly that the tetrameric protein dissociates to monomers during denaturation and that the ligands are not dissociated until denaturation takes place. The enthalpy of denaturation of the protein is 4.57 f 0.25 cal g-’ and is independent of temperature. The enthalpies of dissociation of IPTG and ONPF a t the denaturation temperature are very large, 37 and 42 kcal (mol of ligand)-’, respectively.
T e study of the thermal denaturation of proteins by means of differential scanning calorimetry (DSC)’ has led to many interesting results (Privalov, 1979, 1982). In the case of the reversible denaturation of many small globular proteins such as ribonuclease A (Tsong et al., 1970; Privalov & Khechinashvili, 1974a) and lysozyme (Pfeil & Privalov, 1976), DSC has afforded the most convincing evidence that the process takes place with a surprising degree of cooperativity involving the entire molecule in an all-or-non, or two-state, process, with no intermediate states between the native and unfolded states being significantly populated at equilibrium. The demonstration of two statedness involves the use of the van’t Hoff equation d In K / d T = A H , , / ( R P ) (1) which expresses the variation with temperature of the equilibrium constant, K, for the denaturation process. For any the van’t Hoff, or apreversible two-state reaction, AHVH, parent, enthalpy, must equal AHcal,the calorimetric or true enthalpy. Procedures for evaluating AH,, and AH,,, from DSC data will be outlined below. In the cases of some proteins it has been found that AH,, is significantly smaller than AH,,,. In such cases the denaturation may involve two or more sequential two-state steps or the more or less independent unfolding of two or more separate domains in the protein molecule. In these cases the observed DSC curve, which gives the temperature variation of cex,the apparent excess heat capacity, can be expressed as the composite of appropriate component curves. This decomposition of the DSC curve into component curves assumes reversibility of the process under study. The usual criterion for complete reversibility that is applied in DSC studies is that a reheating of the scanned material after cooling should give a curve that is superimposable on the original curve. On this basis, the denaturation of many proteins turns out to be partially or completely irreversible. +This work was supported by Grants G M 22441 (K.S.M.) and G M 04725 (J.M.S.) from the U S . Public Health Service, C-576 (K.S.M.) from the Robert A. Welch Foundation, and PCM 81-17341 (J.M.S.) from the National Science Foundation. S.P.M. was a NIH predoctoral trainee (GM 07833). * Address correspondence to this author at the Department of Chemistry, Yale University, New Haven, C T 06511.
0006-2960/85/0424-3842$01 SO10
However, this criterion for reversibility may be. too restrictive, as is suggested by the familiar example of the unwinding of double-helical polynucleotides. The thermal denaturation of a polynucleotide such as poly[d(AT)] is fully reversible, demonstrating the fundamental reversibility of the unwinding process. On the other hand, the denaturation of calf thymus DNA appears to be irreversible. This process is rendered reversible if a very few cross-links are introduced between the two chains (Becker et al., 1964). These cross-links serve to prevent actual chain separation with loss of register between the chains. It is thus apparent that the denaturation of the DNA is energetically reversible but that the renaturation is extremely slow. This is further indicated by partial renaturation by annealing at a temperature slightly below the denaturation temperature range. The lactose repressor protein regulates the transcription of the lac metabolizing enzymes in Escherichia coli by specific interaction at the operator site in the genome (Miller & Reznikoff, 1980). This binding is modulated by the presence of sugar ligands bound to the tetrameric protein of M , 150000; inducers decrease the affinity of the protein for operator DNA, while antiinducers stabilize the repressor-operator complex (Miller & Reznikoff, 1980). Two domains within the repressor, amino terminus and core protein, can be isolated by mild proteolytic digestion (Platt et al., 1973). The amino terminus binds to nonspecific DNA and operator-containing DNA (Jovin et al., 1977; Ogata & Gilbert, 1979), while the core protein binds to inducer and to operator DNA (Platt et al., 1973; Matthews, 1979). In this paper we report DSC measurements of the thermal denaturation of the chymotryptic core protein of the lac repressor of E . coli. This reaction appears to be irreversible by the usual DSC criterion in that the denaturational endotherm is completely missing from a rescan of the denatured protein, which appears to have been rendered completely insoluble by the denaturation process. We shall show here that this denaturation, nevertheless, appears to follow quantitatively, both in the absence and in the presence of various ligands, the dictates of equilibrium thermodynamics as expressed in the
’
Abbreviations: DSC, differential scanning calorimetry; IPTG, isopropyl 0-D-thiogalactoside; ONPF, o-nitrophenyl 8-o-fucoside; SDS, sodium dodecyl sulfate; Tris, tris(hydroxymethy1)aminomethane.
0 1985 American Chemical Society
VOL. 24,
DENATURATION OF THE PROTEIN OF LAC REPRESSOR
Table I: Experimental Thermodynamic Data for the Thermal Denaturation of Core Protein protein concn Ah 1" (cal g-l) (kcal mol-') (mg mL-I) tm ("C) 72 1 5.55 1.35 63.80 5.71 742 1.43 63.70 621 4.78 2.44 64.95 58 1 4.47 3.13 64.97 559 4.30 5.74 65.10 580 4.46 11.5 67.00 4.31 560 15.3 66.65 mean 4.80 623 SEb f0.26 *34
NO.
AH"H(1) (kcal mo1-l)" 414 432 532 517 545 518 571 504 *26
1 5 , 1985
3843
Cmax
(cal K-'g-I) 1.130 1.214 1.243 1.129 1.145 1.115 1.190 1.167 *0.022
"AHVH(1)= 9.01RTm2cma,/Ah(see text). bSE, standard error.
van't Hoff equation. We take this as evidence that the denaturation is fundamentally reversible under the conditions of the DSC experiments but that we do not know how operationally to achieve renaturation.
1.2c
RESULTSAND DISCUSSION Denaturation of Core Protein in the Absence of Ligands. The thermal denaturation of core protein was studied at concentrations ranging from 1.35 to 15.3 mg mL-'. The obAfter the work reported here was completed, it was found that no exotherm appears when the protein is scanned at 1 K min? in the new version of the Privalov DSC, the DASM-4. In this instrument the cells are helices of small-bore platinum tubing with horizontal axes. The absence of the exotherm shows that it is in some unexplained way due to the pillbox form of the cells in the earlier instrument and is not due to heat evolution by the protein. This conclusion is supported by the fact that the exotherms when observed did not appear to reduce significantly the denaturational endothermic enthalpies.
1
1
IC 1.0 0
I
MATERIALS AND METHODS Isolation of Lactose Repressor. Repressor was isolated from Escherichia coli CSH 46 according to the methods described by Rosenberg et al. (1977), as modified by O'Gorman et al. (1980). The isolated repressor was 98% pure by SDS gel electrophoresis. Assays were executed as described previously (O'Gorman et al., 1980). Isolation of Core Protein. The core repressor was purified as described previously (Matthews, 1979; Hsieh & Matthews, 1981), with digestion of the intact repressor by chymotrypsin to produce the core domain consisting of amino acid residues 56-360. The isolated core protein exhibited inducer-sensitive operator DNA binding. The concentration of the core protein was determined by absorbance at 280 nm = 0.51). The isolated domain was concentrated to 3-15 mg/mL in a ProDi-Con vacuum concentrator/dialyzer. Following concentration, the preparation was dialyzed for 48-72 h against the buffer utilized in the DSC experiments: 0.048 M potassium phosphate, pH 7.4, 15% glycerol, 0.1 mM dithiothreitol. The protein solutions were frozen and shipped from Houston to New Haven in dry ice. Calorimetry. All scanning calorimetric experiments were performed with a DASM-1 M microcalorimeter (Privalov et al., 1975). A large exotherm was observed during the denaturational endotherm when a scan rate of 1 K min-I was employed, making it impossible to obtain quantitative data concerning the denaturation. It was found empirically that the exotherm, for completely unknown reasons, was reduced in amplitude by reducing the scan rate and by employing relatively high protein concentrations.* All the data reported here were obtained at a scan rate of 0.23 K m i d . A tracing of a typical DSC curve for core protein is given in Figure 1. This curve illustrates the noise level observed in our experiments and the vestigial exotherm at 60 "C.
'
Y
0
c
a W
I
u
k V
W
n ln ln W V X W
TEMPERATURE,
O C
FIGURE 1 : Tracing of a typical DSC scan of core protein, showing
the noise level typical of the scans obtained in this work. Protein concentration was 5.74 mg mL-'; scan rate was 0.23 K m i d . The dashed curve is the best fit of a theoretical curve for the process N4 G 4D to the experimental curve. See text for the method of arriving at this curve.
served results are listed in Table I. Column 1 gives the total protein concentration and column 2 the temperature, t,, at which the excess specific heat, cox,reaches its maximal value, c,,,. The specific enthalpies in column 4 were obtained by planimeter integration of the DSC curves, and those in column 5 by multiplying by the molecular weight of the tetrameric protein, 130000. Column 7 gives the values of c,,,. The procedure for evaluating the van't Hoff enthalpies in column 5 will be given below. All the DSC curves showed the pronounced asymmetry evident in Figure 1. This asymmetry is to be expected if the denaturation is of the form N, mD (2) where N and D represent native and denaturated protein, respectively, with the protein remaining oligomeric up to the denaturational temperature and then undergoing simultaneous denaturation and dissociation. As shown in an earlier paper (Takahashi & Sturtevant, 1981), the excess specific heat, cex, is given by the expression a(1 - a) AhAHvH (3) Cex = m - ( m - 1)a RTZ where a is the extent of denaturation and Ah is the specific enthalpy in caI g-l. Takahashi & Sturtevant (1981) outlined a procedure for fitting the theoretical curve, eq 3, to the experimental data by varying the adjustable parameter TI/*(the
3844 B I 0C H E M I S T R Y
MANLY, MATTHEWS, AND STURTEVANT
Table 11: Thermodynamic Data for the Thermal Denaturation of Core Protein Derived by Curve Fitting protein concn (mg mL-') 1.35 1.43 2.44 3.13 5.74 11.5 15.3
Ah AHp9I AH"u(2) SD (cal gl) (kcal GI-') (kcal mol'l)" (% G"* 5.55 722 40 1 7.7 5.18 673 47 1 7.4 4.45 579 556 6.8 4.1 1 534 555 8.8 4.20 546 550 6.4 4.27 555 559 9.1 4.20 546 580 7.3 mean 4.57 594 525 7.6 SE' f0.25 f32 f30 f0.4 aAHvH(2)is obtained by minimization of the standard deviation of experimental data from a theoretical curve (see text). bSD, standard deviation.
